 Right, I think we're ready to go. And I'm Richard Portes, Professor of Economics at London Business School, and vice chair of the advisory scientific committee, and therefore partly responsible for the organization of this event, but also co-chair of the joint expert group on shadow banking. And I'll come back to that a bit later. I will start by just giving little bios of our panelists and then say a word about what they're going to talk about and then make a few remarks of my own and try to keep it within limits. In alphabetical order, then, we have, and indeed, no, it's not the order on which they're seated. That's OK. That's OK. At the far end, to my right, is Russ Irani, who is assistant professor at the University of Illinois, a CEPR Research affiliate, I'm pleased to say, and took his PhD at NYU Stern and has four papers already in top finance journals and a new working paper on shadow banking. Next to Russ is Catherine Lubosinski, who is professor at the University of Paris II, an independent director of LCH ClearNet, a major clearing organization, the major clearing organization in Europe, has been on the board of the French regulator for banks and insurance, and is a former president of Swerve organization, which some of you will know. So next to Kathleen, closest to me, is Sujit Khabadiyah. He's the home team now. He's the head of the market-based finance division in the ECB's director general of macro policy and financial stability. He was head of research at the Bank of England before that for four years, took his PhD in Oxford, and has a substantial range of publications in good journals on macro-prudential issues on bank regulation, on financial networks, and on contagion, subject here to my heart. This diversity of our panelists reflects the diversity in a sense of the shadow banking sector of our topic, the topic for this session. And Russ will talk about shadow banks and their relation to the US syndicated loan market. Sujit will talk about non-banks. I'll come back to the distinction if there be any between shadow banks and non-banks. Non-banks in the credit cycle and the pro or counter-sixical behavior of non-banks. And Kathleen will talk about central counterparties. This is an amazing panel. The biggest brains on the planet, except for maybe mine. There's no need to listen to them before you vote, before you vote them through. Yeah, right? There's not one old white male among them. They're going to put on a truly great performance. They're going to make Europe great again. We were talking about the problems of Europe. And this is all very exciting. But first, you have to listen to me for a few minutes. Sorry about that. First on shadow banking and non-banks and all that. Some of you will know that the FSB decided, and it's wisdom, that they would replace the term shadow banking by non-bank financial intermediation. And that will come into all their publications, as well as the discourse. This institution has yet to make a decision on that. I think it's very unwise. The reason, supposedly, is that shadow banking carries a pejorative negative connotation. And non-bank financial intermediation, that's a good thing, of course. And since we have now, supposedly, progressed to a resilient system of market-based finance, we should put this shadow banking term back in its box. I think that's wrong. I think we have not progressed to a resilient market-based finance system. There are weaknesses in the system, in the institutions, and the activities that we comprehend in shadow banking. And on that, I suggest that you have a look at our shadow banking monitor, which just came out a couple of weeks ago, which was mentioned yesterday, which I think is an excellent publication, if I do say so, which indicates what the weaknesses and the risks and so forth are. If you just look, and we had a discussion in the general board of ESRB yesterday about the risks, we always do every session. And you can see that there are risks that the risks in asset management, for various reasons, may be going up, that leveraging in hedge funds appears to be going up, that maturity transformation in asset management appears to be going up, that lending to hedge funds is still pro-cyclical. This does not leave me entirely happy or calm, shall we say. So I think that to eliminate the term shadow banking, because we've got it sorted, in a sense, is unwise. The second remark that I would make is on whether shadow banks are systemically important. You can tell a story about insurance companies, for example, saying they're not systemically important, really. If one goes down, well, so what? Well, of course, there is an example from the crisis. You will remember AIG, yes? That's an insurance company that clearly had to be bailed out. And so it depends on what the insurance company is doing and what its interconnections are with the rest of the financial system. And that's the key thing. It's the interconnections, and in particular, between banks and shadow banks. I've done together with some colleagues here and in other central banks some work on the interconnection between banks and shadow banks using EBA data. We'll be able to do more when we get the AIFMD data on stream. And I think there are reasons for concern there. Mario Draghi yesterday talked about the need for macro-approved policies for non-banks. I think he did say non-banks, I'm afraid. But in another sentence, he did talk about the shadow banking monitor. So that's OK. Why do we need macro-pro policies for non-banks or shadow banks? Because there are issues there. And there are issues that are not going to go away for a while. And finally, I would try to be a little bit provocative about the naming question. I don't think it's a trivial question. I'll be a little bit provocative about CCPs, knowing that Kathleen is going to talk in much more detail about them. But CCPs are very interesting. Financial market utility, they've been called. It's part of the infrastructure. And it is a utility. Saying that it's a utility suggests to you, should these really be private sector organizations? Should they be profit-making organizations? And if you think about it a little more, you say, wait a minute here. It's clearly very obvious network economy, if you like, and economies of scale here. Indeed, this is a natural monopoly in principle. It should be. It is. And therefore, it seems to me that, given the interconnections between the existing CCPs, which you will see from Kathleen's talk, given that there are weaknesses of individual CCPs. And we saw an example of that only a few weeks ago in Sweden when a CCP got into trouble and came very close to failing, the so-called waterfall of bringing in capital was almost dry. What would have been done then? Well, that's a very interesting question because there is no resolution procedure in place. And so there are dangers here. A, B, there is a rationale, I think, economic rationale for putting them all under one roof. And that is to say, for consolidating them and having a single CCP. But that I leave for discussion. So I think I'll stop there and let this amazing panel get to work. And first, we'll have Russ Irani. Is that OK? OK, so first of all, let me thank Richard for the invitation. I'm delighted to be here. I'm going to be talking about shadow banks or non-banks in a very narrow context. So I think my talk is going to be complementary with Sujits. His is taking a much more macro perspective. My talk is going to be much more micro in nature. So I'm going to be focusing on the US context and on the US syndicated corporate loan market in particular. So what I'm going to do is I'm first going to provide some basic facts about shadow bank entry into this market. Well, first I'm going to tell you what a shadow bank is in this context. And then I'm going to talk about some basic facts about shadow bank entry in this market. Then I will run through some causes and potential consequences. At least on the academic side, there is a much better understanding of the potential causes of entry in this market. Much less is understood about the potential costs and benefits. And then finally, I'm going to wrap up by talking about some gaps in our data and the related gaps in understanding. So from an academic standpoint, if I wanted to do more research in this area to better understand the costs and the benefits, how you guys could potentially fill in the gaps. So just first to settle on some definitions, so we're all on the same page. I'm going to be talking about non-banks that are unaffiliated with bank holding companies. So many of the shadow banks that people will think about will either have direct connections to bank holding companies through direct organizational affiliations or through guarantees. I'm going to be focusing on unaffiliated institutions. These institutions are going to fund corporate loans with their balance sheet. So they are, for instance, it will be an asset manager that holds a corporate loan in the balance sheet, as opposed to distributing it immediately after purchase. So these institutions might involve leverage. They might involve liquidity and maturity transformation. This is a much more broad definition. Hopefully, we'll get a bit more narrow towards the end. Syndicated loan is a loan that's co-funded by a group of banks and or non-banks. And all of the evidence I'm going to present today is going to be based on a supervisory credit register of syndicated loans called the Shared National Credit Program. This is administered by the US federal regulators, so the FDIC, the Fed, the OCC, and so recently the Office of Thrift Supervision. They've been surveying these syndicated loans going back to the mid-70s, 1977. And what's really interesting about this survey is that it includes a list per loan of all lenders that are funding a given corporate loan, so in one of these syndicates. And it includes all of the banks. That was the primary focus of the regulators. But it also includes all of the non-banks. So this setting gives us a nice laboratory to study how non-banks have been entering this market over several credit cycles. We can see in the distant past and in the current state of effect. So let me just quickly describe what this market looks like. This is 92 to present. Not the levels of the market, but this is showing the composition of the funding in this market. So if you ignore the colors for now and just focus on the trend, if we go from the left-hand bar is 1992, the far right is something like 2014, so more or less present. There has been a shift in the composition of funding from traditional deposit-taking financial institutions, so bank holding companies. In 1992, they were funding about 75% of the market, so the complement that gets us from 25% to 100% will be these traditional banks. And if you fast forward to present, that relationship has flipped. So now you have the lion's share of the market being taken up by non-bank financial institutions that are unaffiliated with both domestic and foreign bank holding companies. The second pattern you're going to see here is there is a shift in the diversity of institutions funding these loans. So you have corporate loan securitization. On the one hand, but you also have asset managers, you have hedge funds, you have private equity, and so on. Those non-bank financial institutions holding these loans in their balance sheets. If you look at the distressed loan segment of the market, the same patterns exist. So in the distant past, it was mostly banks funding this market, funding the loans in this market. And if you fast forward to present, it's now almost entirely funded by non-bank financial institutions. So nearly 100% and we're seeing more asset managers and hedge funds and so on showing up in this market, but also more corporate loan securitization as well. Okay, so why is this going on? So there are two prevailing views here. One is that these shadow banks, these non-bank financial institutions have comparative advantages, so they can provide cheaper funding. They may be bringing new technology to the market as well, which is how they provide this cheaper funding. So if you think about the residential mortgage market, as a leading example, you have Fintech, the way at which they're able to price risks and so on is perhaps better than traditional financial institutions, but also we can think about securitization as bringing lower cost of funding to this market. The other prevailing view is that this is regulatory arbitrage. Okay, so as traditional financial institutions face binding regulatory constraints, so we can think about capital requirements. More recently, we can think about liquidity requirements and so on. There may be, this term that's been used so far, there's more leakages from the traditional sector to these non-bank financial institutions. So in the US, macro-pru has recently shown up in this market. It's specifically in the syndicated loan market, also called the leverage loan market. The Fed issued guidance on underwriting standards, and as a consequence, more of the origination and funding of these loans actually shifted into the non-bank sector. So as a result of this regulation, we observe more leakages to either affiliated entities within the bank holding company. So the leading example here is you have a bank that has a CLO in-house and credit is being shifted into the CLO, which may have lower capital requirements, so swapping $100 of loan into $90 of AAA-rated security and carving out the rest and selling it off. The loans may be funded by unaffiliated entities, or there may be some kind of mix, so you may have an unaffiliated entity plus some guarantee providing the funding. So much more is understood on the causes at this point, as far as the academic side is concerned, there's, I have some work myself looking at capital regulation and capital requirements. For example, the US implementation of Basel III and how that led to leakages from the traditional deposit taking financial institutions towards some of these non-bank financial institutions, much less is understood about the consequences of this shift in credit, mostly because there are gaps in the data, so it's very difficult to actually measure how much of the credit in this particular market is taken up by these non-bank financial institutions. That's why we're leveraging this supervisory data. It's been quite an exercise putting that together, but we believe this is gonna provide opportunities to study in more detail some of the potential consequences. For instance, on the positive side, the supply of credit and also the efficient allocation of risks. On the potential costs of this entry, it's not obvious. Some of the problems that we have, opacity is information asymmetry, which can be due to adverse selection of moral hazard issues. There's mixed evidence documenting how pervasive those issues are, but these non-bank financial institutions certainly have less access to government guarantees and central bank liquidity. So they're potentially an unstable source of funding. I think that's a relatively uncontroversial statement. And so in terms of consequences, this can matter for credit cycles. There could be too much credit in good times, but there could be sudden stops in bad times. As far as the banks that are operating in this market, as you might have gleaned from the first couple of pictures, they're really in the underwriting business. So they're originating and mostly distributing to these non-banks. So that process can break down when the non-bank funding dries up. So this is more recently been referred to as pipeline risk. This can potentially have real effects for borrowers. There can be knock-on effects to the real economy. In the process of trying to document those types of effects in an incredible way. I'm going to show you a couple of pictures on secondary markets. The idea here is that if the non-bank funding is unstable, if it's more slippery, then in the bad state of the world, these non-bank institutions could be scrambling for funding. This could lead them to dump the assets on the market, potentially could lead to a fire cell. On the other hand, if these non-banks are fast-moving sources of capital, we might expect to see capital inflows to at least some of these non-bank institutions and potentially they could be liquidity providers in this market. So these are unresolved questions. I've got a couple of pictures, which, at least in this context, can show how the evidence is very mixed even at this stage. This is showing secondary market trading activity. So the US-indicated loan market, there is an active secondary market for these corporate loans. Participations in these corporate loans are traded over the counter. This is showing, again, 92 to present. We're able to back out trading activity from our credit register by looking at flows. We can actually look at flows across sectors and also flows within loans. And what I'm showing here is on the buy side. So institutions that are flowing into the loan market are picking up participations. And so there is trading activity during normal times that trading activity tends to pick up during bad times. This is actually the distressed segment of the market. Okay, so that's very important. What we see in 2008 and beyond that these non-bank institutions, the trading volatility actually spikes and it stays a lot higher than it was in the past. There's a broad range of non-bank financial institutions showing up in the market. So there are asset managers, there are hedge funds entering the market. Someone surprisingly, it's surprising to me, at least when you first started to do the analysis, you have corporate loan securitization. So CLOs as an active buyer in the distressed segment of the market. So at least at a first glance, this looks like these non-bank institutions or at least some of these non-bank institutions are providing liquidity to the market during these times of stress. And in particular, during the segment of the market where we would expect the traditional financial institutions that may be capital constrained to be offloading credit. Okay, so if you think these banks are attempting to recapitalize through secondary markets, then it appears that these non-banks are actually providing. At least some of these non-banks are providing liquidity. That was quantities, this is prices. Okay, so as I said, there's an active secondary market for participations in this market. This is 2006 through 2011, looking at prices in the market. So at the beginning of the period 2006, these loans were all trading at 100 cents on the dollar. There was a big drop off in the market starting in the middle of 2007, which persisted into 2008 and there was a recovery beyond. So what we do in, this is just some very coarse evidence that's consistent with one of the hypothesis we test in our research. So we're looking at what we can do with this register as we're able to characterize loan syndicates going into the financial crisis. So if you think of the beginning of 2007, the end of 2006, you can, for a given loan which is being traded in this market, you can characterize exactly which institutions are funding that loan. And if we believe that the identity or the liability structure of those institutions, those non-banks funding the loan is gonna matter for fire sales and how that loan is gonna perform in secondary markets during the bad state of the world, you might expect to see some differences in prices showing up. Okay, and so that's exactly what we're documenting in our research and exactly what we're showing in this chart here. So there are two lines here. There's a dark line and there's a dashed line. The roughly speaking, the dark line we can think of as loans where the marginal investor is something like an open-ended loan mutual fund, corporate loan mutual fund, where the liabilities tend to be more short-term, more liquid. And the dashed line is something like an insurance, the marginal investor is something like an insurance company. So long-term stable funding, okay? And we're starting to see some variation across the secondary market prices depending on the liability structure of that non-bank. So you might ask whether or not fire sales matter for every fire seller. There's a fire buyer that might be consistent what we showed in the previous picture, but this is the data point that I want you to consider. Okay, so just to wrap up. So we have some gaps in our data, gaps in our understanding as far as the understanding is concerned. There are open questions here. Okay, we'd like to be able to understand the relative importance of these various factors for explaining this non-bank entry. And then we'd like to also understand in terms of the consequences, we'd like to precisely measure the costs and precisely measure the benefits. We've done nothing along those lines so far in this discussion and also very little in my own research, but that's kind of a direction we'd like to head in. How can we get there? Well, there are various measurement issues and data issues at play here. So in what I think is a fairly remarkable dataset that we're working with, there's a very coarse classification of these shadow banks. There are only coarse entity types. We have no information on how these institutions are funded, say the liability structure of these institutions. We have no idea of how leveled these institutions are, maturity mismatch at illiquidity, whether they're guaranteed and so on. So one leading complication with our analysis is the SEC, the Securities and Exchange Commission, where many of these funds are registered, they do not speak with the Federal Reserve in ways that we hope they would. Okay, all right, thank you. Well done keeping the time too, thank you very much. Kathleen. Oh, so we're not respecting the alphabetical order. That's okay, fine. Are you sure? Oh, sorry, no, I'm sorry, you're absolutely right. You didn't sit down on the record. Sujit, Sujit, Sujit. No, I should have. Stick to plan, never deviate from plan, vote it through, sorry. So I obviously failed the first test of this conference. I'd like to thank Richard and thank the organizers for inviting me to speak here today. Obviously what I'm going to say are my own views and not those of the ECB or of the Euro system. Let me set the scene by just highlighting the importance of market-based finance. I think the President spoke to this in his opening remarks yesterday, and I think he actually spoke to some of the underlying data that's in the left-hand chart, which shows that by March 2018, the total share of the financial sector assets held outside the banking sector, so by non-banks, stood at around 55%. And what you see there is an rapid growth in the non-bank financial sector since the late 1990s across all segments. So our data qualities improve, but you see rapid growth in particular in the investment fund sector, but also elsewhere in the non-bank financial sector. What the right-hand chart focuses on is investment fund holdings of debt security. So it's a rather specific dimension, but I think it's also useful to draw out that quite a lot of this financing is effectively supporting the real economy. So if you look at the right-hand bars in that right-hand chart, what you see is that a large share of non-financial corporate debt, so ultimately real economy financing, is held by investment funds, and that share has grown from approximately 15% to 20% in 2009 to around close to 30% now. Now, part of that reflects the role that investment funds of the non-banks played, acting as a spare tire, as the banking sector recovered after the crisis, but part of it is obviously structural as well. Now, as Richard has said, there's a lot of work that's been done on different dimensions of what is called shadow banking, what I'm going to call market-based finance, and a range of risks have been identified by really important and useful work by the ESRB shadow banking group, by the FSB group that also works in a similar space, by the work of authorities like AOPPA and ESMA, obviously work that's done here and by other national authorities, national central banks, national supervisors, and so on. And maybe just before I go on to the main theme of my talk, it's just useful to recap some of the risks that these reports have highlighted in the non-bank financial sector into three broad categories, interconnectedness, liquidity, and solvency and leverage. Very briefly, on the interconnectedness side, these institutions, these intermediaries are very highly connected or can be very highly connected with the banking sector on both the asset and the liability side. There are also common asset holdings that might be common to, for example, banks, insurance companies, and investment funds. But perhaps one thing that's a bit less recognised is that there are also ownership links. So quite a large proportion of asset managers are effectively owned by either banks or insurance companies, which can pose a reputational risk if the asset managers that are owned by banks get into distress. Liquidity risk has been talked about a lot, obviously there's redemption risk at funds, insurance companies run significant maturity mismatch, which exposes them to interest rate risk, and all of these dimensions can lead to amplification of market liquidity shocks. And obviously leverage can amplify all of these risks, whether that's actual leverage or synthetic leverage. What I want to focus on there today is that I think there's less emphasis on the interplay between market-based finance and the wider credit cycle and the real economy. And this was also a point that I think Philip Blaine made in his keynote remarks yesterday. So what I'm going to spend the next 10, 12 minutes covering are three areas in which I see interplays between market-based finance and the credit cycle. So firstly, the supply of wholesale funding from non-banks to the banking sector. Second, direct financing of the real economy. And third, the interplay with terms and conditions in wholesale markets. There are, I should emphasise, also links to the wider debate here on the global financial cycle, but I'm not going to cover those in the interests of time. Then for illustrative purposes, I'm going to just cover very briefly two pieces of ECB research, which I think are illustrative of the type of work we need to do at the micro level to better understand proselyticality in the non-bank financial system. And I'll conclude with some policy thoughts and open issues. So let's start with wholesale funding. What this chart shows is based on UK data and it shows the growth of debt in the UK from the late 1980s until 2011. Now, when you talk commonly about debt in the UK or generally there's a lot of discussion about before the crisis of growth in household debt and corporate debt, which did increase to some extent. But what's quite striking is that between 2003 and 2007, two-thirds of the growth of debt in the UK was the growth of financial corporate debt to the green stuff in this chart. Largely reflective of intra-financial system activity and a lot of that reflecting the financing of the banking sector by non-bank financial institutions. And it's quite remarkable in terms of the growth in debt. You also see a very sharp cyclicality which this is a more granular focus on repo market activity. This time in the left-hand chart taking US data and the right-hand chart euro area data. And what you see is cycles in repo financing in the US and in the euro area. Again, the idea here is that the non-bank financial sector through its financing of the banking sector is essentially or was prior to the last crisis contributing to proselyticality in the banking sector and ultimately creating wider systemic risks. So that's one key dimension, the non-bank financing of the banking sector which also obviously leads to interconnections. Second dimension, just direct financing. So what this chart, very interesting chart from a recent FSB report shows in red is the non-bank credit cycle. So the blue is the bank credit cycle based on a cross country sample. You see very clear cycles in bank credit. What's striking is that in the most recent the pre-crisis upswing, so in the early 2000s, the red, the non-banks was very highly correlated with the blue. So you see non-bank credit cycles moving in sync with credit cycles and you see that also during the crisis. Now you don't see that throughout history and there's, I think this is a very interesting chart to prompt us to think why this relationship might change over time. But what is evident is that non-banks can also play a clear role in the wider credit cycle through direct financing which is roughly what the red line in this chart reflects. Let me take a specific example of that which again Philip Lane referred to yesterday. He mentioned real estate financing from non-banks and what this chart shows here is zooming in on the insurance sector specifically within the euro area. What you see here is in some countries insurance companies have gained real estate exposure either directly or indirectly through funds and you see that in the Netherlands, in Belgium, in Austria and Cyprus that the extent of these exposures are quite significant and what I think is also quite striking is that in the three countries with the highest residential exposure, there's the ones in yellow, the Netherlands, Belgium and Austria, there are also three countries that received a warning from the ESRB on residential real estate vulnerabilities in 2016. Now, there's not necessarily anything causal in that but what it does suggest is that maybe insurance companies when there's a real estate boom going on might be an amplifier. We also have some information on, for example, commercial real estate bond holdings and what you see in that sector is that investment funds and insurance companies are increasingly supporting commercial real estate lending as well. Last I mentioned is the terms and conditions of transactions in financial markets and I'm here thinking about margins and haircuts. And what you see here is a chart based on a particular hedge fund source which shows that prior to the crisis, if you look at the blue line, there was a compression in the margins. You can't, I mean, given the scale of what happens afterwards, it's not hugely evident but margins compressed prior to the crisis. So essentially the terms and conditions of wholesale transactions got looser and then obviously you had a huge blowout in the crisis but I think this is a really important dimension when thinking about again the interplays between non-banks and the credit cycle because a lot of these terms and conditions are set by the way in which all of the different types of counter parties which do include banks obviously but also a range of non-banks interact in wholesale financial markets and what you see again is clear evidence of prosyclicality in this dimension. Okay, so I said I wanted to show you a couple of examples of some ECB research which speaks to prosyclicality. This is much more micro level, it's much more illustrative and what I want to do, the first set of information is a bunch of aggregate macro type information. These two illustrative studies are just going to some of the types of behaviors that I think might be important for further work for the future. So in one way you can think about this as setting up a future research agenda in this space and these are some initial things that have been done by colleagues here. First one focuses on the investment funds sector and in particular the flow performance relationship and the interplay with leverage. So the flow performance relationship essentially says that when performance of funds is good, there are inflows, when performance is bad, there are outflows, that's well known. But what about the interplay with leverage? Obviously there's already been discussion today and there was yesterday about leverage in the investment fund sector. The left hand chart just sets the scene. I think again some of these figures were alluded to yesterday where you see that there is some leverage but not huge amounts of leverage in aggregate across the hedge fund sector based on an IOSCO survey. But what's quite striking from the right hand chart which is again illustrative because it's purely based on Dutch data which is data that colleagues here worked with the DMB on this particular study. You see a tail of very highly leveraged hedge funds and bond funds. And it's this tail I think that we should be particularly worried about. So what did this study do? It took that Dutch hedge fund data over, sorry, the Dutch investment fund data over the past 10 years or so and it looked at the flow performance relationship. And it looked at it in terms of positive performance and negative performance. So the left hand chart shows after positive pass returns, which you see on the x-axis, there were flows into both leveraged and unleveraged funds which you see on the y-axis. So the flows on the y-axis are a percentage of the lagged assets. Consistent with the theory. But no difference between leveraged and unleveraged funds. What is striking on the right hand chart? When there were negative out turns for funds, so you see that again on the x-axis, you see a striking difference in the outflows from leveraged funds which you see in yellow from the unleveraged funds which you see in blue. So essentially the flow performance relationship was much stronger for leveraged funds. So an interplay here between leveraged and prosyclicality in the fund sector. Second thing I want to show is a study that's been looking at how insurers react to different types of shocks. And there's a theoretical model here which essentially explores the way in which changes in the, if there's a fall in the risk-free rate, then insurers might act in a counter-cyclical manner. But the theory also works on the idea that if there's a decrease in the risk premium, the value of equity of insurers increase, so in those circumstances insurers might act in a prosyclical manner. So that the way in which insurers respond to shocks might depend on the type of shock. So you can think of a boom as one on which risk-premium fall and that would predict prosyclical behavior on insurers. Obviously in a low interest rate environment that would predict more counter-cyclical behavior. So what this study did was looked at how the risk-premier and the risk-free rates evolved and looked at when there were episodes of prosyclical behavior. And what you find is that while there's a lot of counter-cyclical behavior of insurance companies in the sample, there are episodes of prosyclical behavior which both encompass selling, which you see in the yellow, which is when the risk-premier are rising, and buying, which is what you see in the gray bars when risk-premier are falling. So let me wrap up then with a couple of policy thoughts on where this all leaves us. I think the first thing to say, and I think the president said this and I think it was picked up in the press this morning, is we need better data. And this has already been mentioned by Russ in his talk. And I think that is really key in this area. But let me touch on the three areas that I spoke to. On the wholesale funding dependence of banks, progress has been made internationally under Basel III through the establishment of the NSFR and the LCR. But one might also think about the need for time-varying liquidity ratios for banks or, for example, core funding ratios has been used in New Zealand. Macro-prudential instruments for banks, non-banks, however, are much less developed. And again, as the president said yesterday, this is a key area of focus and further work. Areas that I think are important would be to operationalize the macro-prudential leverage limits for AIS, which do exist under the relevant directive. Critically, developing globally comparable leverage measures for investment funds, which is essentially the FSB recommendation, IOSCO, which is essentially asking for better, consistent measures of leverage, which again speak to the data point. Now, that's a really key part of the global debate. Operationalizing liquidity tools for AIS use its additional macro-prudential liquidity tools might also be useful. And then again, as was mentioned by the president yesterday, there's an ongoing debate about macro-prudential instruments for insurers, which I think he mentioned was discussed at the general board yesterday morning, but AOPPA have also published work in this space. Lastly, on terms and conditions in wholesale markets, I think it's important to impose, to have minimum margins and haircut floors on both centrally and non-centrally cleared derivatives and on SFTs. Now, the FSB has agreed to some of this in terms of SFTs, but the previous vice president also spoke about this in a much broader context. In addition, one could also think about time-varying counter-cyclical margins and haircuts as a layer on top of that. Again, I should say in a lot of this space, the FSB has done a lot of work, including publishing a paper I think in 2016 on macro-prudential policy beyond the banking sector. So with that, since I think a minute or two over, let me just leave on the last slide a bunch of open issues that I think are important where we need to do further work on the interplays between the real economy on prosyclicality, leverage and liquidity risk, but critically on the toolkit for the non-bank financial sector. Thanks very much. Thank you, because I can't go around. Okay, well, I'll start like everybody else does. I would like to thank the organizers and Richard for having me invited and use the usual disclaimer about what I will say does not reflect the view of LCH. Being an independent director is not the same as being an independent variable in statistics, but might not also reflect the views of University of Paris too, who doesn't really care about what I'm gonna say today. So, what I've tried to do here is to spot where are the weak links and central clearing functioning in terms of consequences that they might have in systemic risk. Because as Richard pointed out, CCPs are crucial in the functioning of the financial architecture, because all the flows are going to be concentrated there. Now, if we just look at the numbers about the size of CCP measured by assets, remember the non-bank financial intermediation world is about $160 trillion, as estimated by the FSB report, the last one. CCPs belong to the subsectors of OFI, called OFI, or the Financial Institution, which accounts for 99 trillions. Now, when you measure the CCP size by their assets, then they account for 0.4 trillion, which would tend to lead us to think that they're quite negligible, they're very small in size, but there are of systemic importance because they offer those critical services. Now, so if you look in terms of individual size, they're important because it's the same issues when you think in terms of concentration, so the CCP sector is highly concentrated. There's also lots of, sorry, a bit. Lots of interconnectedness that's involved and complexity. Now, if we look at the concentration aspect, just to give you a rough idea, you have one CCP that clears about $1 trillion a day, $1 trillion a day, so the flows, in terms of flows, they are critical. Now, concentration is still a trend in the financial industry. I mean, concentration, you have it in the banking sector, of course, more or less depending on which part of the world you're looking at, but it's a trend in the asset management business. If you look at the first three asset managers, they have $10,000 billion under management. So if you combine this with an increased share of, or an increasing share of passive management strategies, one can understand the consequences you'd have in terms of financial stability when one of these major asset manager decides to sell a given public debt for a given country. They can be destabilizing. If we look at the concentration, do you have it here a little bit? Yeah, well, if you look at the concentration trend for CCPs, you have a fabulous report from FBIS, FSB, CPIM, from August 18 on the central clearing and through dependencies. You'll have all the information you want, but you see that the two largest CCPs in terms of pre-funded financial resources account for 40% of the total pre-funded resources for the CCPs. So for some product, as Richard pointed out, they are quasi-monopolies. For instance, in Europe, interest rate swaps are cleared at 90% through only one CCP. That cleared last year $850 trillion in notional. So of course, this results from the increasing level of clearing derivatives, OTC derivatives. And if you compare the level of clearing in 2009, only less than a quarter of OTC derivatives were cleared, where now we're approaching interest rates swaps, now we're approaching 62%. So in the US, if you look at in the US, you'd have more or less the same thing, that is the CME that clears most futures contract, OCC clears most of the option contracts, OIS clears most of the CDSs. So there's lots of concentration and quasi-monopolies in some instances. Now, you can understand why this concentration occurs. The functioning of netting positions is subject as to economies of scales and to network effects. That's obvious. So now, if we look at the interconnection, well, post-crisis regulation achieved to mitigate some interconnectness between banks, mainly with the obligation to clear OTC derivatives in CCPs. Okay, and again, remember interest rate swaps for those who are not too familiar. There are about $300 trillion of notional, even if they've come down. Now, the drawback that comes from this obligation of clearing is that the interconnection between banks and CCP has increased to a lesser extent, the interconnection between banks and asset management, as asset managers has also increased. Lots of reason, as you know, about one would be, for instance, the gross of synthetic ETFs, which increases the link between asset managers and investment banks, or the fact that asset managers don't always have a direct access to CCPs and they clear through banks, so interconnections. So it seems that when you don't measure only with the asset side, interconnection has someone shifted from within the banking sector to banks, non-banks, right? Yeah, that's fine. If you look at some other numbers, what you would see is that, for instance, one CCP clears for 29 systemic financial institutions. Seven CCP's clear for at least 16 globally systemic financial institutions. So you have those interconnections between CCP's and CFI's that have increased. There's been a reallocation of interconnectedness, but propagation and amplification channels are still in the financial system and we should still be worried. Another way of looking at it is saying that, oh my, I'm lost, sorry, 11 of the largest clearing members out of more than 300, let's say, are connected to between 16 and 25 CCP channels. Okay, and when I say large clearing member, we use the measure, the size, by pre-funded financial resources. So what happens when one of these largest clearing member, which is bound to be a CFI defaults, well, it will probably default in all the other CCP's too. Even, yeah, it will probably. So you see the propagation mechanism is there. Then the second problem is that there are amplification loops, clearing members, or very often, members that are service providers. They provide custody, custody services, settlement services, credit and liquidity facilities, cash and non-investment cash. So if a member, if a big member defaults and is also a service provider, then the CCP is impacted both ways, yeah, so it's even worse. So another problem is linked to the prosyclicality of margins and haircuts. Now, you know that most collateral is either cash or government bonds, so any sovereign debt crisis has a big impact on margin calls. Article 41 of EMEA set out requirement for CCP's to monitor the prosyclicality arising from margin revisions and margin parameters, so CCP's have to adopt at least one of the three anti-prosyclicality margin measures related to volatility estimates. Because volatility estimates are key to prevent erosion of margin levels during quiet periods, which would generate a short prosyclical correction where markets condition to change. So the three choices in volatility estimates are either you take a 10-year look back period, which is a bit concerning now, because we're 10 years away from the big financial crisis or the big mess of September or October 2018. So what, 208, sorry. Now some exchanges have decided to just drop that crisis scenario because we're moving on and some other exchanges will still keep it, so see the risk is not managed the same. The other two choices they have in volatility estimates is either put a 25 weight to stress observation in the look back period or put margin buffers of at least 25% of calculated margin. So here you introduce discrepancies between CCP's and we had some issues in CCP's all around CCP's after the Italian elections when the two-year government bone spread surge for more than 200 basis point. You can imagine that you wanted to increase some margins. Oh I forgot to show you a beautiful graph that I really liked in that report I mentioned. That's the graph of interdependencies between CCP's which are in red and clearing members that provide services. So again it comes from the same report I mentioned and I think it's this graph is better than a long speech. Now questions for policy makers because Tom is going. So and then I will go back to Richard when he says you know what's the optimal shareholder structure because they tend to be monopolies, they're utilities and recognize as such they're too important to fail. Definitely they should be public entities. Well if you go on you know if you read Darryl Duffy he says there should be only one CCP around the world. So I mean it's not you know it's not obvious. Of course you could wonder if they should be totally private or if they should be owned by their users like DTCC is and DTCC is still efficient and but what you notice is you have either silo structure or horizontal structure. The silo structure is preferred as for example in Germany Eurex is known clear CCP is owned by Eurex. Usually it's the markets that own the CCP with one exception in the US being DTCC and in Europe being LCH where some users are also shareholders but that is really an issue. Anyway there are conflict of interest everywhere and we know that public entities are not always the most efficient ones as we have learned in France with the Crédit Lyonnais default which was a public bank. So being public is not always optimal right? What I think policy makers should focus on is to how to mitigate the interconnection between clearing members and service providers. Okay like CCP should deal only with CSDs or have credit line with central banks and not with other banks because if a clearing bank is a member and that same clearing bank provides you with credit line and that bank defaults what are you going to do? You can't use those credit line anymore. So and that's why some CCP's are very happy to well in some ways are happy to have the status of the bank so some of them are banks like LCH I say well weird banks and then because they're banks they have direct access and direct liquidity access with the central bank otherwise you have to negotiate some agreement with the central bank but we have a Brexit issue here for instance what's gonna happen when the Bank of England which has guaranteed some liquidity in euros to British CCP's what happens when the Brexit occurs do they still have those swaps agreement between central banks? I mean we don't know, we don't want to know and then we have so and along the lines should governments be granting more licenses or foster consolidation on CCP's because there are economies of scale because risk management is very sophisticated and you need lots of IT inputs and so it costs a lot and we do see that small CCP's are more see more risky than others. For example you mentioned one CCP which everybody knows it was a newspaper it was NASDAQ electricity with the problem that occurred almost failed well they had as a clearing member an independent trader yes very rich but an independent trader which in the end turned out not to be rich enough. Well this could not occur like in France where you're not allowed to have individuals being clearing members okay so those are issues to be looking at and then there's another it's not equivalence or interoperability what I meant is in the process of giving equivalence licenses then you're introducing an uneven playing field simply because CCP's are not managed the same way and some CCP's haircuts have a floor which is linked to the ECB floor on haircuts and some CCP's the haircuts are much smaller than those used by the central bank so it appears more risky and then you have a choice of a system where you have preferred CCP within a trade or you introduce interoperability between CCP's. Interoperability seems to be a trend because it's more flexible but I would say looking at it from the systemic point of view it increases propagation because if one CCP defaults and has links with is interoperable with five other CCP's than the other five other CCP's are in a situation okay so there are lots more, lots more question we could ask but I think I'm running out of time sorry thank you. Well that's all three of you have put a number of issues and I wonder I think we ought to have a little discussion among us and then turn it over to the floor and I'll just raise a couple of questions for my colleagues here but they can raise questions for each other. Russ you said that there were some arguments for the shadow banks taking on more and more of the syndicated loan business that they're better, they're cheaper, well yes but are they better at evaluating loans, do they have the expertise that banks supposedly at least have in making lending decisions and then monitoring loans exposed and if I think of if I generalize that point I think about the rising direct lending by private equity firms which I think poses potential dangers aside from just piling leverage on leverage as somebody has put it it's again a question as to whether those firms have the expertise to evaluate to evaluate the loans. Sujit you pointed out in the beginning the ownership links between banks and substantial parts of the shadow banking system especially insurance companies and asset managers too. Do you think that poses any particular dangers that regulators should get themselves involved with? One aspect of the interconnectedness that we see between the shadow banking sector and the bank sector. Kathleen I hadn't seen this paper by Darrell Duffy. I'm delighted to hear he agrees with me does he? Well it was a single case. He was studying the optimal number of CCPs and he said well one would be. Optimal number of CCPs is one exactly. Well well it must not default huh? I think one is a simple number, it's a simple number right but it also, it does mean if you want to call it socializing it it's a public utility, what's wrong with that? I think the Crédit Lyonnais example is a complete, it just is not appropriate. This CCP would not have Jean-Yves Abarrère as its CEO entrusted by Mitterrand with the objective of making Crédit Lyonnais a wonderful global investment bank. You know you can't, I'm not in favor necessarily of socializing things. We have this little issue in the United Kingdom some of you will realize about re-nationalization of some of the sectors that were privatized. But in this case it seems to me that there is an argument for making this a public utility with direct involvement by the state or by a collective of states and of course with their serving as the lender of last resort. So here we go. Starters for 10 as they say. But maybe we have a couple of other questions from the panelists to each other. Russ? Yeah, just to follow up with Suji I had a couple of questions. So there's this great slide you had up with the various risks that are showing up from the different non-market-based finance like leverage, liquidity and so on. And I was wondering, well I have a couple of questions. I was wondering on that point do you have a sense on which factors are most important or is it even possible to kind of get a sense of which factors are most relevant because of course that's going to feed into the policy design and the policy response. And then maybe just another question which was kind of a shorter question I thought that the performance flow relationship and how it ties into leverage was very interesting. But I didn't get a sense of why that relationship was showing up the way it was so why leverage would amplify that maybe there's something simple that I didn't quite catch. Kathleen, any questions for the other two? Kathleen, yeah. Oh yeah, well I could say well some people have showed that competition is good for the innovation process. Not for answers yet, just questions. We'll get to answers in a minute. Oh okay, so I'm just focusing on my answers, sorry. Yeah, but fine, I have no question. Sujit. I think one observation I find very interesting on Catherine's presentation was this 10-year aspect which it came up in some meetings that we were having in London last week and I think I did find that quite a concern that there's this potential that essentially a big episode of stress drops out of models and therefore you essentially run the risk of kind of re-amplifying making the presentiality of margins worse again just because of the sort of aspects where an episode that we know was pretty problematic drops out of the models and it just seemed to me that that was an interesting thing that I'd also picked up on and whether it might be appropriate to sort of replace three years of data with those periods that were 2007 or 2008 and then the other thing that actually struck me was whether this is just an issue for CCPs where it's obviously tied to the stretch of regulation but whether this might be a more, this is not just about shadow banking now or non-banks but whether there's a wider issue here as that period, if that period drops out of models and maybe that's something that we should be collectively thinking about a bit more. Why don't you continue and answer if you wish. Yeah, I can. So on this issue of which we see is the most important risks. I mean so in the financial stability review we've highlighted liquidity risk in the investment fund sector as a particular amplifier so I think that is something that we are particularly concerned about. What I have a sense of though recently is that there is increasing discussion about potential for rising leverage as well in pockets of the sector so I think where we think about it is that liquidity risk I think we know is a major concern because as one market participant put it to me recently before the subprime, so in the subprime crisis one we're thinking about is that a bunch of very low quality assets were packaged up into a security that was AAA. And there is a question, this is what one market participant put to me was is the fund industry or parts of the fund industry now basically packaging up a bunch of very illiquid assets that all have a very low underlying liquidity and essentially selling them as a liquid promise via an ETF let's say. And I think that is, and the analogy there I think was quite striking and I think that's something that we're quite, you know, is of concern and the liquidity risk is definitely a concern. What I would say on leverage I think we are quite constrained by the data so in an aggregate picture as I showed you in one of my slides it doesn't look like the fund sector is particularly highly leveraged but there's a tale of funds. We only see euro area domiciled funds so what we don't know is are there pockets of leverage that might be occurring outside the area in funds that are domiciled on the Cayman Islands and synthetic leverage so I think there really, again as I said before the IOSCA work on consistent measures of leverage is really important that that gets delivered on and I think we need to get a better picture so I wouldn't necessarily say we have definitely a big concern about leverage but I think we could well be concerned about it but it's just hard to gauge but for me that would be a very significant risk and there is some indicative evidence of some concerns in the tale. I think in terms of the amplifiers I mean obviously what I was showing was what the data speak that have been done from Spurts and some colleagues I think it's just it's essentially more of a concern that if a fund is highly leveraged it just is very susceptible also basically the sentiment in that fund is more susceptible to the fact that basically it's not all equity ultimately and therefore this is just a great susceptibility and propensity that the other investors might wish to pull out because they're worried about the overall potential I guess the health of the fund in an overall sense. Sorry I should have mentioned on the ownership structure I mean again relative to the liquidity and the leverage sorry I'm just jumping backwards actually I would say liquidity leverage are key concerns the interconnectedness is a concern the ownership structure I think is something worth monitoring I would say if you were not to meet a bit of relative sense on it I would probably say I'd be more still concerned about the liquidity and the leverage and the direct connections and the common holdings partly because some of the largest asset managers in the world are independently run as well it was just an observation that I think it's you know there is an ownership structure they obviously are completely separate in one sense but there might be a potential for a reputational risk and you see this in other aspects and I think it's something that I guess this I would say is more probably a micro potential sort of issue to the supervisors of those particular institutions where there is a common ownership structure should be just cognizant of this dimension in the way in which those individual insurance companies and banks that deal in asset managers are run You will recall that right at the beginning of the crisis right at the beginning one of the events was that a couple of banks had to pick up their liabilities effectively of their funds yeah that was a bit supposedly independent yeah it was a bit different there were contingent lines that were written so I think it's obviously important to understand in the structures as they are legally set up there should not be that sort of now obviously that's what I'm saying it's important in a supervisory perspective to ensure that those contingent lines and those implicit guarantees that you refer to are not present I think that's an important question any additions before I want to get to the floor okay about re-involving governments in the management of the CCP I'd be a bit skeptical because you would kind of reintroduce the sovereign risk bank risk loop replacing bank by CCP because in case of a crisis you need to change the margin the haircuts and the governments could be tempted to temper that a bit I think what is more important is to put the CCP's on a more level playing field so you have to harmonize the regulation around the world I'm not sure small CCP's are equipped to do all the investment necessary to really properly run the risks and do the risk analysis and so maybe granting license well I mean it's complicated I mean it's an issue that people should investigate and to answer your question about the 10 year look back I wouldn't be too worried because what you can always do is in your stress scenarios that you use you can still keep it like some institution or still keeping the 1987 scenario in a stress scenario so it's not you can keep it the ability to do it is obviously there it's where the people actually do it but then that's what the supervisors have to check in terms of risk management or do they all you know run a scenario you know it's a very interesting point about the monitoring and so on I think in the RNBS market so in the US there's been a lot of work talking about the foreclosure rates of those loans and it seems to be the case that loans that were sold by banks and securitized had higher foreclosure rates as opposed to the rate of which those loans are restructured which is presumably bad for the homeowner in the case of the corporate loan market you have this process of syndication before securitization where the norms in the market are for the bank arranging the loan to actually retain some exposure and therefore to deal with the monitoring and so on of the loan and the renegotiation in the case of some kind of problem with the borrower so it's possible that the market itself has mechanisms in place to deal with this particular problem so that if you have more funding from outside of the traditional banking sector you can still have renegotiation and so on taking place in an efficient manner So let's open the floor to questions, comments, objections, whatever Marco, Marco Pagano Have we got a microphone? No, he's right there One question that has not been addressed by the panel as far as I can see is the issue of solvency and resolution of CCPs So essentially, you know, Catherine said that it's important for CCPs to have access to central bank liquidity and therefore maybe to be managed by bank or an institution which has the status of a bank so that kind of addresses the issue of liquidity but you know what if the overall resources of the CCPs anyway are not sufficient and so this opens the issue of bailouts and it kind of intersects with the other issue that you were also raising about ownership structure rather than ownership structure whether you want to have a single CCP or several CCPs because if you have a single CCP in the world which government should step in and kind of bail it out so I think these are non-trivial issues so there is a kind of tradeoff between efficiency and you know how you manage bailout Let's take a couple more Is there any? Yeah, that gentleman there, yes, first Blue shirt, yes Hi, I'm Stephen Moore from the SSM Just wondering, is there any research being carried out yet in terms of the impact of shadow banking that you've talked about there in terms of lending but also the shadow banking in terms of fintechs in the way that they are eroding the profitability of existing banks and from a supervisory point of view we look at how profitable banks are to generate capital to be resilient in downturn events and yes, I can't see you but I know you're there Bernard Winkler, ECB So I want to come back to Richard to your opening so I agree with you we should not get rid of the term shadow banks but I would take an opposite conclusion maybe we should get rid of shadow banks properly defined Now Lehman, ten years after that crisis was not a crisis of traditional banking nor was it a crisis of traditional market finance it was a crisis of the hybrid and there I think rightly called shadow banking quasi-bank-like activities which mixed the two pure models of finance So I get very nervous there were a lot of financial intermediaries who would love to get access to liquidity from the central bank or guarantees or this or that so we need to be very clear what is the public policy rationale So central banks are the land of last resort to banks which are deposit taking institutions they are maturity transformation they are leveraged and they do monitoring of their loans on their balance sheet, that's banking that's the public policy case for us being involved So if we start to worry about asset managers liquidity is that a shadow bank? I would be very very hesitant to go to that direction meaning we have to market potentially sort of underwrite or sort of counteract their risk taking we made a big mistake already with the money market funds because the public perceived a money market fund to be a quasi-deposit but it was not breaking the buck was possible so that's not promised too much anybody investing in equity and investing in a fund takes the risk including the liquidity risk on the day it wants to redeem So I want to really draw a sharp line here because there's a slippery slope now of going in the direction we want to market potentially sort of supervise everything and this at the end ends up on the public sector balance sheet one more point the broker-dealer sector the first thing that happened at the core of the financial crisis was they were transformed into bank holding companies within weeks after Lehman to have access to the central bank balance sheet the discount window Do we want to have that kind of dynamic in the next crisis? Thank you very much Well, you would have seen both Goldman and Morgan Stanley fold there's CDS spreads that got up to 2,500 3,000 basis points One more and then we'll come to the panel Yeah, Ola Feiglin Yes, I'll be I was also having a question on the CCPs I mean one of the issues is that access to CCPs and becoming a clear member is sort of expensive and therefore many institutions small institutions clear indirectly via a clearing member in Europe this works such that the clear member takes on a principal relationship so I'm wondering maybe also for those who are feeding a bit sentimental about the possible loss of shadow banking should we consider and look a bit more into shadow CCPs? Okay, I think that gives us enough to be going on for a little while Who wants to start off? I don't know I'll start by answering Marco Pagano's question about recovery and resolution you know the CCPs now have recovery and resolution plans The problem, okay we have to keep in mind we don't have time to talk about that but they have to comply with what's called the covered two scenario and have enough profunded resources in case of two big the biggest and one other clearing member's default to be on the safe side what you can do is just look what happens when the two biggest members clearing members default are your funds default fund big enough liquidity etc so when properly managed you see that a CCP can easily survive to a covered two default what worries me is that in this scenario if it's the two biggest clearing members that means there are two globally sci-fi's so if you have two defaults from globally systemic banks that also have accounts in several many CCP's I think we're in a situation there because probably there's more than two members that will default and more than one CCP that will default and then you know anything can happen because in the end the closure of a service and etc before it closed service you have to call non-funded resources from the members but if they're in a situation they can't provide those even if they're committed those liquidity lines so you know so it's a problem with the recovery and resolution scenario because of this huge interconnectedness but if just one or two and nothing else happens anywhere else I mean there's no problem it's all in a properly managed CCP it's all set but it's the interactions now okay well time is going by so maybe you want to answer this the others later okay thanks so let's pick up on a couple of questions I won't come up on the CCP issue it's been well covered maybe actually on terminology I think one way to think about this is you know shadow banking was associated with the term when it first arose was that takes many different forms it's bond markets it's insurance companies it's investment funds and you know for me I think just whether it's called market based finance or non-bank intermediation I think that captures in a broad sense the sorts of things we're worried about so I think you know for me that's the key is that I think I think it was clear from my talk I don't think we're sanguine about the risk just because the name is the name is changed by the FSB I think we are very concerned about those risks I think that was also clear again in Phillip Lane and the president's remarks yesterday so I wouldn't personally be concerned that the terminology changes meaning that we're all sanguine about the risks so the question about you know also how far you extend their macro-represential policy across the sector I mean for me I think again my emphasis on credits and credit cycles was also partly because I think we should be thinking here so what we want to avoid is essentially the ex-post scenario that you're talking about where there is a distress and where you might need intervention ultimately and if we think about these intermediaries as increasingly financing the real economy contributing to credit booms which there's good evidence that credit booms often tend sour what we need to think about I think is ex-ante policies and then that might well be ex-ante macro-represential instruments applied potentially to funds to insurers and so on and then the last thing I think there was a question on shallow banking and fintech which I think was actually a talk by Amit Seru at a conference at the ECB and he has a paper which I think is very interesting but there's a literature on this but it eventually finds that I think in the U.S. mortgage market that there are certain brokers that basically you can type in information and get a mortgage in not seconds minutes let's say there are essentially eroding the margins of the traditional banking sector and that's in a very specific study but I think there's also some evidence from the Nordic banks of this going on in terms of increasing competition for new players in certain Nordic countries so I think that is a really important issue and I didn't touch on it but I fully agree that's something to think about where there is some emerging research Yeah just to follow up on that at the same point there is emerging I think the Seru piece on the U.S. residential mortgage market does show that at least the online lenders in the U.S. are coming in but they're targeting very specific segments of the market so typically a bank will originate alone and then one of these online lenders will come on and refinance at the kind of top end of the market where the loan can then subsequently be sold to a GSE these institutions don't have balance sheets they just be a pass through straight to a GSE Okay I'll finish up we have one minute with the broad ranging very deep and broad ranging issues that were raised which cannot be dealt with in one minute but just a couple of quick comments one is yes we may need regulation of these entities and I give you an example the asset management sector when I discuss this with people in that sector they often say but look you know what's the problem you don't want to give us capital regulation we don't have balance sheets excuse me all the risk is taken by our investors what's the big deal here if something bad happens then it's spread out among goodness knows how many investors no problem well not quite because there's one obvious risk and that is firefail risk that an asset manager in trouble redemptions rise etc etc ask to sell assets the big ones have big positions they have to have big positions you can't be you can't hold you can't be running in assets without having big positions in the markets and if you have to liquidate big positions that then has ripple effects throughout the financial system so I don't think we can ignore that the question is to whether ultimately you may have to bail one or another out again imagine what would have happened if Morgan Stanley and Goldman had not been allowed to become bank holding companies if AIG had not been bailed out I don't want to let my imagination go that far really I wouldn't like to see it I wouldn't like to see the consequences so the consequences of that are that we are out of time and it remains only for all of you to thank our three outstanding panelists who have made your upgrade again thank you